Esta coleção reúne imagens ópticas adquiridas pelos sensores WFI (Wide Field Imager), embarcados nos satélites do INPE: CBERS-4, CBERS-4A e AMAZONIA-1. As imagens referem-se às regiões de Franca (SP), Franca Nordeste (SPNE) e partes da Amazônia brasileira, com destaque para o monitoramento de áreas naturais, queimadas, cobertura de nuvens e corpos hídricos. Os arquivos seguem um padrão de nomenclatura que codifica informações como a localidade (ex.: Franca_AMAZONIA, Franca_SP), o sensor utilizado, a data da imagem (AAAAMMDD), órbita / ponto, o nível de processamento (ex.: L4), composições espectrais (ex.: BAND4321, BAND16151413) e tipo de processamento (ex.: RegAWFI). Os dados estão disponíveis nos formatos .tif e .png.

CBERS-4/MUX - Level-4 Surface Reflectance product over Brazil and part of South America. L4 SR product provides orthorectified surface reflectance images. This dataset is provided as Cloud Optimized GeoTIFF (COG).

CBERS4/WFI - Level-2 Digital Number product. Level 2 products have radiometric correction and geometric correction using satellite ephemeris and attitude data (system correction).

CBERS-4/PAN10M - Level-4 Digital Number. L4 product provides orthorectified images.

AMAZONIA-1/WFI - Level-2 Digital Number product. Level 2 products have radiometric correction and geometric correction using satellite ephemeris and attitude data (system correction).

The MODIS-Aqua Monthly Remote Sensing Reflectance (Rrs, unit sr-1) provides 8 spectral bands temporal resolution of one month and spatial resolution of 1 km over the Brazil oceanic waters and open ocean South Atlantic waters. This collection captures 7 visible, and 1 infrared channels using Level-1A images acquired from the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument onboard the NASA's Aqua satellite. The Level-1A data were processed into Level-1B and GEO using the Data Processing Tools (modis_L1B) from SeaWiFS Data Analysis System (SeaDAS) software. The Level-1B and GEO data were applied in the atmospheric correction OC-SMART (Fan et al., 2021) by the National Institute for Space Research (INPE, Brazil) and the Laboratoire d'Océanologie et de Géosciences (LOG, France) generating the Level-2 data. The Level-2 data has been mosaicked to generate daily maps capturing the complete Brazilian ocean waters. The daily mosaic data were reprojected to geographical (lat/lon) coordinates using as reference the European Space Agency (ESA) Ocean Colour - Climate Change Initiative (OC-CCI) into Level-3 grid. Both, the mosaic and the reprojection were done using the Sentinel Application Platform (SNAP) on its version 10. Finally, the daily reprojected data were temporal merged to create the monthly Rrs products using the geometric mean.

Sentinel-2 image mosaic of Brazilian Cerrado biome with 10m of spatial resolution. The mosaic was prepared in support of TerraClass project. The false color composition is based on the MSI bands 8, 4 and 3 assigned to RGB channels. The temporal composition encompasses 04-months of images, starting in june 2022 and ending in September 2022, with a best pixel selection approach called Least Cloud Cover First (LCF). More information on LCF can be found at Brazil Data Cube web site (https://brazil-data-cube.github.io/specifications/processing-flow.html#temporal-compositing). This Image Mosaic used more than 14000 Sentinel-2 scenes and was generated based on an existing data cube of Sentinel-2 images.

Sentinel-2 image mosaic of Brazilian Yanomami Indigenous Territory with 10m of spatial resolution. The mosaic was prepared to support the partnership between the INPE's Health Information Investigation Laboratory (LiSS) and the ICIT-FIOCRUZ Health Information Laboratory (LIS) a multi-institutional body coordinated by Fiocruz and the ministry of health, by creating a health situation database of the Yanomami Indigenous Land. The false color composition is based on the MSI bands 11, 8A and 4 assigned to RGB channels. The temporal composition encompasses 06-months of images, starting in April 2019 and ending in September 2022, with a best pixel selection approach called Least Cloud Cover First (LCF). More information on LCF can be found at Brazil Data Cube web site (https://brazil-data-cube.github.io/specifications/processing-flow.html#temporal-compositing). This Image Mosaic used more than 15000 Sentinel-2 scenes and was generated based on an existing data cube of Sentinel-2 images.

This is a land cover classification map of Brazilian Cerrado, from August 29th 2017 to August 29th 2018. This classification was made on top of Landsat-8 days data cubes with spatial resolution of 30 meters, using the best pixel composition function named as Least Cloud Cover First (LCF), which was previously named Stack in BDC older versions. The input datacube was Landsat-8 - OLI - Cube Stack 16 days - v001, which was deprecated. The classification model was trained using 48850 sample points spread across the Cerrado biome (Annual Crop 6887, Cerradao 4211, Cerrado 16251, Natural Non Vegetation 38, Open_Cerrado 5658, Pasture 12894, Perennial Crop 68, Silviculture 805, Sugarcane 1775, Water 263). The spectral band used were B1, B2, B3, B4, B5, B6, and B7 along with the vegetation indices EVI and NDVI; the clouded observation were identified using the Fmask4 algorithm and estimated using linear interpolation. The classification algorithm was TempCNN (Deep Learning). This product was funded by the Brazilian Development Bank (BNDES).

CBERS-4A/WFI image mosaic of Brazil Paraíba State with 55m of spatial resolution. The mosaic was prepared in order to demonstrate the technological capabilities of the Brazil Data Cube project tools. The false color composition is based on the WFI bands 16, 15 and 14 assigned to RGB channels. The temporal composition encompasses 03-months of images, starting in July 2020 and ending in September 2020, with a best pixel selection approach called Least Cloud Cover First (LCF). More information on LCF can be found at Brazil Data Cube web site (https://brazil-data-cube.github.io/specifications/processing-flow.html#temporal-compositing). This Image Mosaic used more than 50 CBERS-4A scenes and was generated based on an existing CBERS-4A/WFI image collection.